D
dgmoore
All,
I have deployed several similar but distinct web sites on our company
intranet to disseminate schedule status and related information on
large projects (thousands of tasks, hundreds of users). The data for
these sites is maintained by teams of schedule analysts in an Oracle
database in one instance, in SQL Server in another instance, and in
MSProject in two additional instances.
The web site content consists of a highly sructured collection of task
detail pages (all text) and a large variety of Gantt charts (detail
schedules, summary schedules, schedules sorted/filtered by functional
area, by organization, and various other sorts as well). The Gantts are
generated from Project using EditCopyPicture, and are imagemapped to
provide drilldown from summary schedules to detailed schedules, and
from detailed schedules to milestone detail pages. The pages on which
the Gantts are displayed also present a rich variety of other related
project information taken from the respective underlying databases,
with extensive hyperlinking to other information on sites external to
the schedule site.
The sites are updated each day in batch mode and serve thousands of
html pages and hundreds of Gantt charts. The process by which the files
are generated is based on VBA through an MSAccess interface to the
respective data sources via ODBC and ADO.
These sites are updated daily; it takes about 10 - 15 minutes every
morning to generate a site update.
These sites work very well, and users are quite happy with them.
However, the process of producing thousands of html and gif files every
day and moving them to their respective web serves is a little
unwieldy, and is inefficient in that it requires replacing pages on the
server regardless of whether the underlying data has changed or not.
I am continuing to seek a more elegant solution - the text content on
these pages could easily be pulled live from their underlying
databases, but I remain stumped on the question of the Gantt charts.
The underlying data requires extensive massaging before it can be
exported to Project to generate Gantts, and I'm at a loss as to how it
would be possible, in response to a browser request, to assemble the
data, produce a Gantt chart, put it on an imagemapped page, and serve
it to the user. Latency is an issue - a 'canned' Gantt chart can be
served in a fraction of a second, but the queries used to produce the
Gantts are very intricate and can take a long time to run, so a large
proportion of the Gantts could take many seconds to generate on the
fly.
I'd welcome any suggestions from the experts on this one!
Thanks
Dave ([email protected])
(Remove the NOSPAM to email)
I have deployed several similar but distinct web sites on our company
intranet to disseminate schedule status and related information on
large projects (thousands of tasks, hundreds of users). The data for
these sites is maintained by teams of schedule analysts in an Oracle
database in one instance, in SQL Server in another instance, and in
MSProject in two additional instances.
The web site content consists of a highly sructured collection of task
detail pages (all text) and a large variety of Gantt charts (detail
schedules, summary schedules, schedules sorted/filtered by functional
area, by organization, and various other sorts as well). The Gantts are
generated from Project using EditCopyPicture, and are imagemapped to
provide drilldown from summary schedules to detailed schedules, and
from detailed schedules to milestone detail pages. The pages on which
the Gantts are displayed also present a rich variety of other related
project information taken from the respective underlying databases,
with extensive hyperlinking to other information on sites external to
the schedule site.
The sites are updated each day in batch mode and serve thousands of
html pages and hundreds of Gantt charts. The process by which the files
are generated is based on VBA through an MSAccess interface to the
respective data sources via ODBC and ADO.
These sites are updated daily; it takes about 10 - 15 minutes every
morning to generate a site update.
These sites work very well, and users are quite happy with them.
However, the process of producing thousands of html and gif files every
day and moving them to their respective web serves is a little
unwieldy, and is inefficient in that it requires replacing pages on the
server regardless of whether the underlying data has changed or not.
I am continuing to seek a more elegant solution - the text content on
these pages could easily be pulled live from their underlying
databases, but I remain stumped on the question of the Gantt charts.
The underlying data requires extensive massaging before it can be
exported to Project to generate Gantts, and I'm at a loss as to how it
would be possible, in response to a browser request, to assemble the
data, produce a Gantt chart, put it on an imagemapped page, and serve
it to the user. Latency is an issue - a 'canned' Gantt chart can be
served in a fraction of a second, but the queries used to produce the
Gantts are very intricate and can take a long time to run, so a large
proportion of the Gantts could take many seconds to generate on the
fly.
I'd welcome any suggestions from the experts on this one!
Thanks
Dave ([email protected])
(Remove the NOSPAM to email)